Measurement-Oriented Multiphysics Forward Modeling and Quantitative Source-Component Analysis of LVH-Related Cardiac Magnetic Fields Using SERF-MCG

Accurate measurement and quantitative interpretation of ultra-weak cardiac magnetic fields remain major challenges in spin-exchange relaxation-free magnetocardiography (SERF-MCG). In left ventricular hypertrophy (LVH), cardiac magnetic field measurements are affected not only by ventricular wall thickening, but also by anisotropic conduction disturbance, fibrosis-related current redistribution, mechanical loading, and mechano-electric feedback. These coupled effects make conventional forward electrical models insufficient for explaining the measured SERF-MCG spatial maps. To address this measurement problem, we propose a physics-informed 3D electro-mechanical-magnetic framework linking representative LVH remodeling to observable SERF-MCG features and enabling separate forward evaluation of predefined myocardial source components. The framework integrates human ventricular cellular electrophysiology, bidomain-based anisotropic electrical propagation, fibrosis-related conduction attenuation, nonlinear myocardial mechanics, and bidirectional mechano-electric feedback. Stretch-activated currents and deformation-dependent conductivity updating are incorporated to describe the mechanical modulation of electrical activity. The resulting myocardial primary currents, volume-conducted currents, and electromechanically modulated equivalent sources are further mapped to body-surface magnetic fields through Biot-Savart-based forward electromagnetic computation. Simulation results show that LVH produces measurable changes in cardiac magnetic field characteristics, including increased magnetic field amplitude, enhanced spatial dispersion, and a systematic shift in bipolar-map deflection angle from 10°-15° under normal conditions to 25°-40° under LVH conditions. Multichannel SERF-MCG measurements and echocardiographic data provide group-level consistency support for the model-predicted LVH phenotypes, demonstrating consistent differences between normal and LVH subjects in magnetic field topology, amplitude distribution, and spatial orientation. Source contribution analysis indicates that myocardial primary currents dominate the measured magnetic field, whereas mechano-electric feedback introduces temporally meaningful modulation, particularly in repolarization-related features. The framework establishes a representative multiphysics forward pathway linking LVH-related remodeling to observable SERF-MCG phenotypes. Group-level experimental comparisons support its ability to reproduce major LVH-related magnetic-field trends, while forward decomposition of predefined source components provides mechanism-oriented interpretation without implying patient-specific inverse localization.